ML20128G018

Forwards SECY-92-339, Evaluation of GE Test Program to Support Design Certification for Simplified Bwr

ML20128G018
Person / Time
Site:	05200004
Issue date:	02/09/1993
From:	Borchardt R Office of Nuclear Reactor Regulation
To:	Marriott P GENERAL ELECTRIC CO.
References
NUDOCS 9302120121
Download: ML20128G018 (2)

Text

____ _ __. _ __

y 1 ,

February 9, 1993 Docket No.52-004 Mr. Patrick W. Marriott, Manager Licensing & Consulting Services GE Nuclear Energy 175 Curtner Avenue San Jose, California 95125

SUBJECT:

TORWARDING OF COMMISSION PAPER COVERING THE NUCLEAR REGULATORY COMMISSION (NRC) STAFF'S EVALUATION OF THE SlHPLiflED BOILING WATER REACTOR (SBWR) TEST PROGRAM Enclosed is a copy of SECY-92-339, " Evaluation of the General Electric Company's (GE's) Test Program to Support Design Certification for the Simpli-fied Boiling Water Reactor " Enclosure 2 of the Commission paper describes the staff's evaluation of the GE SBWR testing program and identifies concerns that GE must address before the staff can complete its design review and issue -

a final design approval for the SBWR. Should you have any questions regarding this matter, please call the project manager, Melinda Malloy, at (301) 504-1178.

Sincerely, OdoissiSfdsdrtyi Richard W. Borchardt, Acting Director Standardization Project Directorate .

Associate Directorate for Advanced Reactors and License Renewal Office of Nuclear Reactor Regulation

Enclosure:

SECY-92-339 dtd 10/06/92 cc w/ enclosure:

See next page -

11002G QLEIRIE!J1103:

. Central; file) PDS1 d/F TMurley/FMiraglia, 12G18 PDR WRussell,12G18 JPartlow, 12G18 TGody, Sr. 12G18 DCrutchfield WTravers JMoore, 15B18 JThompson JNWilson EDO R/F, W9100211 BToms, W9100211 NRR Mailrm, W9100211 PMagnanelli, W9100211 PShea (2)

GGrant, EDO RBorchardt MMalloy ACRS (11) CPoslusn

)

}

0FC: LA:PDST: - R SC:P ADAR (AD) DST:ADAR NAME: PShe - JNW1 RBQrbardt I\

DATE: 02 , 02/ / 02q/93 9302120121 930209 a i

' f PDR ADOCK 05200004 P PDR -

9 1 .

Mr. Patrick W. Marriott Docket No.52-004 General Electric Company cc: Mr. Laurence S. Gifford GE Nuclear Energy 12300 Twinbrook Parkway Suite 315 Rockville, Maryland 20852 Director, Criteria & Standards Division Office of Radiation Programs U. S. Environmental Protection Agency 401 M Street, S.W.

Washington, D.C. 20460 Mr. Daniel F. Giessing U. S. Department of Energy NE-42 Washington, D.C. 20585 Mr. Jeffrey C. Baechler GE Nuclear Energy 175 Curtner Avenue, MC-782 San Jose, California 95123 Mr. Frank A. Ross Program Manager, ALWR Office of LWR Safety & Technology U.S. Department of Energy NE-42 19901 Germantown Road Germantown, Maryland 20874

M M45FM M M M M M M MAulrmurmurmamur i

ps* *%,,

POLICY ISSU5 October 6, 1992 (NEGATIVE CONSENT) SECY-92-339 epi: The Commissioners Er_QD: James M. Taylor Executive Director for Operr.tions Subiect: EVALUATION OF THE GENERAL ELECTRIC COMPANY'S (GE's) TEST PROGRAM TO SUPPORT DESIGN CERTIFICATION FOR THE SIMPLIFIED BOILING WATER REACTOR (SBWR)

Purpose:

To inform the Commission of the staff's evaluation of testing already performed by or for GE, or planned for the future; to identify deficiencies and weaknesses in the test program; and to inforn the Commission of the staff's plans to require GE to address the staff's concerns.

BAgkcround In SECY-91-057, "Early Review of AP600 and SBWR Research Needs," the staff presented its plans for performing an early review of important new safety features of the two advanced, passive light water reactors (LWRs), Westinghouse Electric Corporation's AP600, and GE's Simplified Boiling Water Reactor (SBWR). In SECY-91-239,

" Preapplication Reviews of Advanced LWR Designs," the staff outlined the elements of the process for preapplication evaluations of the vendors' test programs to support passive plant design certification. In SECY-91-273,.

" Review of Vendors' Test Programs to-Support the Design Certification of Passive Light NOTE: TO BE MADE PUBLICLY AVAILABLE WHEN THE FINAL SRM IS MADE AVAILABLE

Contact:

A. Thadani, NRR rru_nna.

& # *;!* ~ *l *'~~

f}

p M- Enclosure

,o u o o u /------ _

• i .

The Commissioners Water Reactors," the staff presented details of the structured process for conducting these reviews. Enclosure 2 to SECY-91-273,

" Issues Requiring Vendor Testing for AP-600 Design Certification," discussed in detail Westinghouse's testing program and the staff's evaluation thereof, and provided recommendations for additional t. sting by the vendor to address deficiencios and weaknesses ,

in the planned program. ~ Enclosure 2 to this paper presents a similar detailed evaluation ,

of GE's test program for SBWR. 1 An advance copy of this paper was provided to the Commission on February 26,.1992. This updated version takes into account comments received from the ACRS on the staff's evaluation of GE's test program, and the results of meetings between the staff and GE between February and the present.

Discussion GE's SBWR depends upon passive systems to accomplish all safety-related functions in the event of a transient or an accident, including emergency core coolant (ECC) injection and reactor decay heat removal.

These processes.are driven either by natural convection, by elevation differences, or by stored energy, such as compressed gas. .The SBWR also includes an automatic depressurization system (ADS), which11s used to reduce the primary system pressure to a-value approximately' equal to that of the containment. This action is essential to allow the pasuive low. pressure ECC injection s,rstems to work properly.- The SBWR design does not utilize any safety-related high-pressure safety injection and depends on early system depressurization to allow for low-head safety. injection. -Valves used in.

the safety systems that are required to=

change state for.those systems to function are driven either by differential pressure.

(check valves) or by battery-supplied de power. Also, many safety system valves in.

the SBWR, including most of those used in-the depressurization system, are " squib" valves, which-rely on a small explosive charge for.

actuation, and cannot be closed once.they;are

, ' - _g

I.

The Commissioners open. All active systems in the SBWR design are nonsafety-related, including pumped coolant injection, residual heat removal, and the diesel generators.

Enclosure 1 contains a brief description and illustrations of the SBWR design.

GE has developed an experimental program to study the behavior of a number of the passive safety components and systems included in the SBWR design. A complementary code development program is also in progress to validato GE's computational models for SBWR analyses. GE has completed most of the -

developmental testing planned specifically for SBWR systems, including (1) Scaled (1:400) low-pressure tests of a heat exchanger in the " GIRAFFE" integral facility at Toshiba in Japan, in both isolation condenser (IC) and passive containment cooling system (PCCS) modes; (2) Scaled (18508) low-pressure tests in the

" GIST" facility at GE, of the gravity drain cooling system (GDCS);

(3) Squib valve experimentation at Wyle Laboratories, including tests of the explosive charges after thermal and radiation exposures simulating those in the plant environment, and tests of the operation of full-scale valves.

Testing in progress includes condensation heat transfer experiments at MIT and the University of Ca'ifornia, Berkeley, for the purpose of developing correlati'ons for use in modeling the performance of the IC and PCCS.

In addition to using results from the SBWR test program, GE is depending on considerable use of technology and experience from development of the evolutionary ABWR, including the advanced controls, fine motion control rod drives, and modular construction.

GE considers the developmental test program to comprise the only " required" testing

L The Commissioners needed to satisfy design certification requirements. However, GE is planning to conduct " confirmatory" tests of safety systems and components, as well. -These tests consist oft (1) Full-scale IC and PCCS heat exchanger component (separate effects) tests, to be performed in the " PANTHER 5" test facility at SIET in Italy in 1992-93, and (2) Scaled (1:25) low-pressure integral tests of passive safety systems, including the IC and PCCS, to be performed at the Paul.Scherrer Institute (PSI) in the " PANDA" facility in Switzerland in 1993-95.

The staff conducted a preliminary review of GE's test program, and requested that GE provide details of the test. plans.for completed and planned work. 'The staff-requested that GE provide'the qualified raw data, as defined in SECY-91-273, from those tests that have been completed. The staff will review these data.

Based upon the staff's initial evaluation, a-number of concerns have been identified that must be addressed satisfactorily by GE. In-Enclosure 2, the staff discusses the results of the review of the test program, identifying weaknesses and. deficiencies, the staff also presents its recommendations for additional testing to address these concerns.

.The issue of " required" versus " confirmatory" vendor-sponsored testing is also examined.

The-staff informed GE on November 6, 1991, of-its preliminary concerns regarding the-SBWR testing program, and discussed its evaluation of GE's test program at several meetings between December 6, 1991,-and July 20, 1992.

The staff also presented its evaluation to the ACRS Subcommittee on Thermal-Hydraulic Phenomena in meetings on April 23, 1992, and.

June 2, 1992, and to the full ACRS on May 7,

'1992, and June 4, 1992.

~

< j i

.l! -..

v i

'f The Commissioners' -

5-The ACRS, in its June 10, 1992, letter to Chairman Selin, essentially endorsed the .

staff's evaluation of the vendor's test-program, and recommended-additional testing l and data assessment. The ACRS-letter is included as Enclosure 3 to this paper. i Egnplusions: Unless otherwise instructed by the-Commission, the staff wilit_

l

1. Transmit to GE the findings _in this_  ;

paper, presented in Enclosure 2,_ 10 days  !

after the date of this_ paper; and

2. Require GE to address the-identified concerns-prior to issuance of Final Design Approval for the SBWR.- ,

t

/

BW aos M.gaylor i

xecutive Director for Operations .

Enclosurest

1. Description of the SBWR Design- ,

'2. Evaluation ofJGE SBWR Testing-program-

3. Letter: dated June 10, 1992, from-D. A. Ward, ACRS, to I. Selin, Chairman,. '

U.S. Nuclear Regulatory Commission, _

Subject:

Testing and Analysis 1 programs  :

in' Support of the Simplified. Boiling. '

Water Reactor Design certification SECY NOTE: In the absence of_ instructions'to'the contrary, SECY-:

will notify the staff on Wednesday, October 21, 1992, that the Commission, by; negative consent, assents to-the-' action proposed in this paper.

- DISTRIBUTION = ,

~ Commissioners-OGC OCAA

- OIG1 EDO.

ACRS SECY , ,

?

4 e ++**r awww- /_ --_sm_m_--n a4- wv w m --e--+ _ u -d w at- - h- w _r- er e- a Mae - '- +

l l

l i

E!1CbOSURE 1 DESCRIPTION OF THE SBWR DESIGli The Simplified Boiling Water Reactor (SBWR) represents a major departure from current-generation BWR designs. It is a " passive" plants all safety injection and safety-related heat removal systems depend on natural convection or gravity-driven coolant flow. Stored energy is also employed in the form of compressed gan or batteries. Valves included in the safety systems are either check valves, which operate as a result of differential pressures, or valves requiring battery-supplied de power that must change state only once and then remain in their final state.

The SBWR design is illustrated in Figs. 1-3. A key feature of the plant is the absence of reactor coolant pumps. Normal reactor coolant flow is supplied by natural convection, driven by the difference in density between the fluid in the downcomer and that in the core and above-core components. An unheated

" chimney" is included to reduce the average density of the fluid above the core, thereby enhancing the natural convection flow.

Because there are no pumps, there are also no recirculation lines, and there are no coolant inlet lines in the reactor vessel below the elevation of the top of the core. The reactor vessel is large in relation to the thermal power of the core, which is only 3/4 of the length of a conventional BWR, resulting in a low core power density. The rest of the system, such as steam separation and drying equipment, reactor control systems, and pressure suppression pool, and including balance of plant, is similar in many respects to current-generation plants, but incorporates advanced technology, such as fine-motion control rod drives, which can be actuated either hydraulically or electrically.

The safety systems of the SBWR are significantly different from conventional BWRs. A natural circulation isolation condenser (IC) is included for decay heat removal up to full reactor pressure. Heat is transferred into large pools of water outside of the containment (the " isolation condenser pools"), which are allowed to boil and exhaust to the environment.

There are no pumped emergency core coolant (ECC) _ systems, and _ no -

safety-related high-pressure injection. An automatic depressurization system is included, consisting of conventional safety / relief valves (SRVs) located on the main steam lines, which are sparged into the pressure suppression pool, and special depressurization valves (DPVs) on separate reactor vessel stub lines and also on the main steam lines, which exhaust directly to the drywell. The DPVs are explosive-actuated " squib" valves. A small explosive charge is Fet off by a battery-supplied de signal, causing a diaphragm to break and open the valve, once the valve is open, it cannot bE YOClosed. In a loss-of-coolant

i ,

l 2

accident (LOCA), the ADS opens the SRVs and DPVs in stages, reducing the system pressure at a rapid rate. Emergency core cooling is supplied by a gravity-drain cooling system (GDCS), the water for which is contained in a large pool in the drywell above the elevation of the core. When the difference between the reactor coolant system (RCS) and drywell pressures falls to a value smaller than the elevation head of the water in the GDCS pool above the RCS, check valves open and allow injection into the RCS through six ECC injection lines.

Heat rejection after a LOCA is also accomplished passively.

Passive containment cooling system (PCCS) heat exchangers, separate from those used for the isolation condensers, are also located in the isolation condenser pools. The inlets to these heat exchangers are always open to the drywell. Steam released to the drywell through the DPVs and the break is condensed in the PCCS heat exchangers; the condensate returns to the GDCS pools for recirculation to the reactor vessel, while non-condensible gases are purgod to the suppression pool air space (wetwell).

In an anticipated transient without scram (ATWS), a standby liquid control system (SLCS) is available to inject a highly

. borated solution into the reactor vessel. The SLCS is supplied by a gas-pressurized accumulator, and is actuated by a squib-type explosive valve. A shutoff valve, controlled by the icvel in the SLCS accumulator, prevents the gas in the accumulator from entering the RCS.

In addition to the safety-related systems, active non-safety systems are available as the "first line of defense" in an accident or transient. These include pumped residual heat removal and coolant injection systems, power for which is provided by nonsafety-related diesel generators.

i

1 ill ,

!  !; l llllll>

ll?lil r

e y

r .

d s n /r e r

e t oo i

t i v

t el a a r r d mu a c r e o a

p n -

i i d c si t _

r l o e r r m l

a a t

e u g e e -

mt at i n -

6 NInF l

e s

s e

V r ,

t o fC j

g" c g j a

e a

' _f d{

j

{

g R l gl iJ g

R .

c' W

B S

ij' 4

SBWR Isolation Condensers Passive Containment Cooling Reactor Isolation 7 _ gg __ 7-m -- s, e,g -~ 1;

$ ,_ll~= $

bhm._=_l; u _ __ %g w  :;

w x n fz m x E i i - W, i i

%n/w/a __at n ,

w

'/////1!////) f

-- - 1 g

• y w i

, ~ d ---

a $

Y/// M/7/

e 2 e a Three 50% units Three 50% units l Passive operation Passive operation l

.SBWR Gravi ty Driven ECCS Short Term Long Term l

( D l-GDCS Pool n

usri 2

[

Reactor .

vessel l 2 -

N >'4 -E l= >< 'd waa c D  ! Th.rm.wu.a.d DekgoVWw.

! til('ll perine-k=ded with fusbie Enk)

!___8***"' W - -___

3 loops

- 11oop = success

h ENCLOSURE 2 ,

Evaluation of GE SBWR Testing Program General Electric's SBWR testing program consists of a series of both separate effects and integral tests. This approach is similar to that employed by Westinghouse for the AP600 design; ,

that test program was evaluated in SECY-91-273 and SECY-92-030.  !

All SBWR integral tests are carried out at a reduced pressure (compared to reactor operating pressure) of approximately 100 psia, about 10 percent of normal operating pressure. .

Despite the similarity of some SBWR components, such as safety / relief valves (SRVs) and the isolation condenser (IC), to those in current plants, operational conditions and the reliance placed on the components for.the SBWR are considerably different from current generation plants. As a result, there are a number of issues that must be resolved involving the performance characteristics of passive safety systems and components. Soma ,

of the key issues include:

(1) Thermal-hydraulic stability.

(2) Behavior of " biased-open" check valves.

(3) Operation of isolation condenser (IC) in presence of non-condensible gases and possible contaminants.

(4) Operation of passive containment cooling system (PCCS) in presence of non-condensible gases and possible contaminants, and operation of non-condensible purging system.

(5) Operation of " squib"-type depressurization valves (DPVs),

both from the thermal-hydraulic viewpoint (blowdown behavior), and also in terms of reliability of operation after-exposure to operating environment.

(6) Operation of plant in "long-term" cooling mode, with steam condensed from containment atmosphere returned to gravity drain cooling system (GDCS) pools and from there to reactor vessel.

(7) Boron distribution in primary system in the event of SLCS operation.

(8) Interactions between safety systems, e.g., IC and squib.

DPVs, which are connected to common stub lines from the reactor vessel.

2 (9) Interactions between safety systems and non-safety, front-line active systems, such as reactor water cleanup (RWCU) system and control rod drive (CRD) injection system.

(10) Severe accident performance tests, as discussed in item 8 on pages 9-12 of this enclosure.

GE's test program appears to provide adequate coverage of some of the above issues. However, there are several areas that are not addressed, or are addressed inadequately. These are discussed below.

1. Thermal-Hydraulic Stability Stability in natural circulation systems is a principal concern.

Power and flow esci11ations can lead to undesirable thermal-hydraulic conditions, such as dryout; thermal-hydraulic upsets can also be exacerbated by oscillatory neutronic feedback in these events. GE has performed stability analysen using the FABLE and TRACG computer codes. These analyses, reported to the staff in a January 1991 meeting between NRC, the Electric Power Research Institute (EPRI), and GE, showed that the SBWR is very

. stable with respect to the density-wave type of oscillation under normal operating conditions. However, the staff expressed concern that GE's modeling did not properly account for the large open chimney at the core exit, and that representative test data would be needed to validate codes used to study the stability of SBWR. In order to further evaluate whether such studies are necessary, an independent assessment of the stability of the SBWR design described to the staff in January 1991 was carried out for the NRC by Oak Ridge National Laboratory (ORNL) . This assessment included calculations with the NRC/ORNL developed LAPUR computer code which showed that, while the system appears to be very stable under normal operating conditions, there are abnormal conditions of operation that might be reached under credible transient sequences that can result in the onset of density-wave power and flow esci11ations. In addition, a low flow and low power instability due to a geysering cffect between parallel channels has been identified as a concern'during normal operating transients such as start-up and shutdown.

On December 6, 1991, the staff not with EPRI and GE to discuss the EPRI/GE response to the staff conclusions that more extensive SBWR stability studies are needed and that codes which have-been validated against thermal-hydraulic tests representative of the SBWR design, including the large open chimney, would be needed to perform these studies. EPRI/GE informed the staff that the chimney design has been changed and that existing experiments are representative of the divided chimney design now employed. GE will validate its codes for density-wave instability studies against these experiments and provide results of the work for NRC review. The geysering instability is being studied using scaled

I i

?

experiments performed by a Japanese partner to GE. The Japanese SATARI code will be validated against these experiments and used for analytical prediction of stable operating boundaries. GE plans to recommend start-up/ shutdown procedures similar to those used in the Dutch Dodewaard reactor to avoid geysering instability. EPRI/GE believes that SBWR is not vulnerable to a loop-type instability also reported by the Japanese; rather, it was characteristic of the atypical experimental apparatus used.

In summary, EPRI/GE have modified the SBWR conceptual design and have identified existing experimental data which they believe to be appropriate for validation of codes to be used for stability studies. They have indicated agreement with the staff that such studies will be needed to confirm the stability behavior of SBWR during various transient scenarios (including ATWS). EPRI/GE cannot today provide sufficient information to permit NRC i ovaluation of the applicability and sufficiency of the foreign .

experiments they have identified for use during code validation.

They have agreed to make this information available to the NRC as soon as they obtain permission from the foreign sources. Until these experiments can be reviewed by NRC, the potential need for additional experiments to support stability evaluation for design certification remains open.

2. Biased-Ocen Check Valves GE has stated that the check valves in the safety systems of the .

SBWR are " biased-open," or held open under conditions of no backpressure. This is apparently in response to a concern that low pressure differentials will not reliably open the valves sufficiently and keep them open. The design specifications for these valves were described to the staff in the January 24, 1992, meeting, but a final design has not yet been chosen. There are two staff concerns with respect to these valvest (a) the pressure drop characteristics of the valves under both forward and backflow conditions, and (b) dynamic response of valve due to pressure fluctuations, especially those that might exist during depressurization. GE has stated that there are currently no plans to test these valves. In view of the' critical importance of the valves to the operation-of the passive safety systems, GE's position does not appear to be adequate. It is recommended that GE develop a comprehencive check-valve testing program addressing the above concerns, or justify the absence _of such testing by submitting data from previous check valve tests that cover check valve designs and the_ range of operational conditions-anticipated in SBWR.

-W* e -

se- e--- - -e nw ,y- -ms---o- yaermy-a- e w ws- a w <---9r-T--N-----FP-~=

l 4

3. IEqlation condenser operatiqD The isolation condenser (IC) serves as the safety-related, full-reactor-pressure decay heat rejection system for SDRR. In operation of the IC, steam from the primary system is condensed, and the heat is rejected to an ex-containment water pool, which is ultimately allowed to boil and is vented to the environment.

Condensate is returned by gravity drain to the reactor vessel.

The IC has provisions for venting non-condensibles to the wetwell. GE states that IC operation is supported by experience gained at current-generation plants, such as Oyster Creek. In addition, a reasonably comprehensive test program is being conducted, consisting of separate-effects heat transfer tests; the GIRAFFE 1:400 scale, full-height, low-pressure integral test; the PANTHERS full-scale component test (separate effects); and the PANDA 1:25 scale, low-pressure integral test. At present, it would appear that this program is sufficient to generate necessary data; however, results from completed tests and test plans for future tests need to be examined to determine if the full pressure range and expected operational conditions have been adequat91y covered in these experiments. This includes heat transfer in the presence of non-condensibles, and possible degradation of heat transfer due to fouling from aerosols that might be entrained with the steam, particularly under severe accident conditions. The existing database for ICs in current-generation plants also should be examined to determine its applicability to this system. Finally, the configuration of the IC system allows it to interact with other safety-related components and systems. For instance, the hot legs of the ICs are connected to the same stub lines as the DPVs. This means that the DPVs can be in communication with the RCS hot and cold legs simultaneously, and both IC and DPV operation may be affected. Also, the heat rejection pools interconnect with those used for the passive containment cooling system (PCCS), which could affect the heat removal capability for both systems. GE should evaluate the potential for systems interactions and show that theco interactions will not interfere with proper operation of the ICs.

4. Passive Containment Coolina System Operation The PCCS is similar in design to the isolation condenser, in that it consiste of vertical, shell-and-tube heat exchangers (HXs) immersed in the same pools as the ICs. However, the tube side of each HX ic in communication with the upper drywell of the SBWR plant. In an event that results in release of steam to the drywell (LOCA or depressurization through the squib blowdown valves), the increased drywell pressure provides the driving force to circulate the stear and non-condensible gases in the drywell through the PCCS, with condensate returned to the GDCS pools and non-condensibles purged to the wetwell. The PCCS takes steam from the drywell, and is expected to operate primarily at i

l l

,. J l

5 1 i

low ambient containment pressures, but will be exposed to the full range of containment conditions. The test program for the PCCS paralicis closely that for the IC system, using the same test facilities. Unlike the IC, however, operating-plant data do not exist for the PCCS. It appears that the program as designed is adequate; however, the test plans and data must be reviewed to l determine if (1) an adequate range of steam-gas mixtures is included, (2) an adequate range of non-condensible gas conditions (composition, pressure, entrained aerosols) is tested, including possible severe accident conditions, (3) an adequate range of heat loads is used (also including possible severe accident conditions and fouling of the PCCS heat exchangers by aerosol deposition), (4) the non-condensible gas purge system works properly, and (5) consideration has been given to the potential carryover of core debris and subsequent rupture of the HX tubes following a severe accident. The possible carry-over of aerosols into the GDCS pools, and the effects of this material on system behavior, should also be evaluated. This program should also address the issue of long-term cooling operation, for which a low-pressure integral test should be adequate.

Both the PCCS and IC heat exchangers will be tested in the PANTHERS test facility at SIET in Piacenza, Italy. Although GE originally classified these tests as " confirmatory," the staff believes that the tests are required for design certification.

GE has agreed to perform these tests on a schedule compatible with that fer final design approval of the SBWR.

5. Ooeration of Exolosive ("Saulb") Deoressyrlzation VaJygs The automatic depressurization system of the SBWR comprises two sets of valves: conventional safety / relief valves (SRVs), which discharge from the main steam lines to the pressure suppression pool, and explosive-actuated squib depressurization valves (DPVs), which are connected to stub lines from the reactor vessel, and discharge directly.to the drywell. GE states that squib valves have a long history of use, and has performed a number of tests designed to show that the valve and its component parts, especially the explosive charges that actuate the valve, will operate reliably over the range of conditions and aging they are expected to experience in the SBWR. While the program included an_ extensive range of tests, there were very few tests of the entire valve assembly under prototypic conditions; prototypic tests (full-size valve) were performed only three times, and details on test conditions have not been made available for staff review, Regardless of the details, however, it can be stated that the small number of tests is insufficient to establish that the DPVs will operate reliably in the reactor environment.

6 The operation of the squib valves is crucial to SBWR safety, particula;1y in the event of a small-break LOCA. Since there are ng safety-related, high-pressure ECC systems, the reactor must be depressurized to allow inventory makeup from the various low-pressure ECC systems. The SRVs may be unable to accomplish the job alone, since they will not depressurize the primary system as rapidly as when combined with the DPVs, nor can they lower the reactor coolant system pressure to that of drywell, sinen their discharge spargers are submerged in the suppression pool. Givon the critical nature of these valves, it appears that further testing of the entire valve assembly under prototypic situations, representing a range of expected thermal-hydraulic and environmental conditions, is necessary to establish that it will operate properly and reliably in the plant. The staff will discuss its views on necessary additional testing with GE.

6. Boron Distribution Durina SLCS Iniection In the event of an ATWS, borated water from the standby liquid control system (SLCS) will be injected into the primary system.

The supply of borated water is maintained in a gas-pressurized accumulator, which injects when a squib-type isolation valve is fired and reactor system drops below the accumulator pressure.

Because of the low flow rates expected when this system-operates, it is important to determine if the boron will diffuse into the core in such a manner as to shut the reactor down quickly enough and with adequate margin. GE asserts that data from ABWR testing will be used to support the SBWR design. However, there are substantial differences in the designs of the ABWR and SBWR ,

SLCSs, including ABWR's use of a pumped, rather than a passive system. GE has reviewed the ABWR test and data, and intends to demonstrate to the NRC that they are adequate for the SBWR.

7. Interactions Amona Safety Systems and Between Safety and Enn-Safety Systems As was true with AP600, a major issue in the design of the SBWR is the potential for interactions among the various safety and non-safety systems in the event of an accident. Although the SBWR safety system design is somewhat less complicated than the AP600 because of the lack of high-pressure injection systems, there is still the potential for interactions. One example of these interactions has been mentioned previoucly since the IC and squib DPVs_are connected to the same stub lines, the DPVs are in communication with the RCS at two points--the stub lines and the IC cold leg inlet nozzle--through this interaction path. It is not clear how the presence of a communication path through the IC will affect blowdown behavior. In addition, operation of the DPVs will initially affect IC operation. Since the removal of decay heat shifts from the ICs_to the PCCS once depressurization occurs, t'.+.is behavior may not pose a problem. However, once the system is fully depretsurized (RCS pressure approximately equal

7 to dryvell pressure), a flow path in addition to that through the PCCS exists for the steam and non-condensible gases in the drywell, back through the DPVs into the RCS and/or the ICs. The effects of this alternate flow path on PCCS system performance should be evaluated. Other interactions that can be postulated include:

(a) Containment backpressure effect on blowdown rate through DPVs, and interactions with PCCS.

(b) Suppression pool pressurization and relief to drywell through vacuum breakers.

(c) Level fluctuation effects during blowdown. Low-low vessel level plus time delays activate various stages of the ADS.

The vessel level will be affected, however, by the opening of the ADS valves. It is unclear how blowdown behavior will be affected by such level changes.

(d) Suppression pool and GDCS in long-term cooling mode.

Gravity-drain ECC injection from the suppression pool (SP) begins about 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br /> after injection from the GDCS pools begins. Once the SP injection path is opened, there are two free surfaces (GDCS pools and SP) at two different elevations, and it is unclear how the two pools might interact. An adverse case would be a " manometer" type of oscillation between the two pools, with little or no net mass addition to the RCS.

(c) The nonsafety-related systems that are available to inject water into the primary system can act in such a way as to change the temperature distribution in the RCS (injection of cold water into the reactor vessel), thereby changing natural convection flow rates. It is not clear if these would be beneficial or detrimental effects.

The preferred way to study interactive behavior is in integral tests, in which system components, configurations, and thermal-hydraulic conditions can be simulated with satisfactory accuracy.

GE has performed limited integral tests and is planning further work in this area. The prior programs have some clear limitations such ast (a) The Toshiba " GIRAFFE" IC/PCCS integral tests were done in a full-height facility, but at a substantially reduced volumetric scale (1:400), which did not represent the current configuration of the systems tested. For example, a single heat exchanger was.used to represent both the IC and the PCCS, while the current design employs separate heat exchangers for these systems. While a 1:400 scale is adequate for some types of integral tests (e.g., SPES for the AP600), the purpose of GIRAFFE was to simulate long-l

h 8

term decay heat removal through the PCCS. However, the heat exchanger configuration in the facility was not representative of tha full-size system, and important effects, such as the distribution of non-condensibles in the heat exchanger header, cannot be characterized due to the one-dimensional nature of the GIRAFFE configuration.

A larger-scale simulation of the systems should be performed to verify extrapolation of these results (heat transfer, non-condensibles behavior).

(b) The GDCS tests performed at GE in the GIST facility did not represent the current configuration of the system. A t single water tank was useo to simulate both the GDCS pool and the suppression pool, while the current design employs separate pools for these systems. Furthermore, the simulated blowdowns and transients were not performed from full reactor pressure. In addition, although this facility was full-height, its volumetric scale was also small (11508).

Tests in the " PANDA" loop, a 1:25 volumetric scale facility at the Paul Scherrer (PSI) Institute in Switzerland, are planned by

- . GE in the 1993-95 time frame. The PANDA facility was described at a meeting between the staff and GE on January 24, 1992. It will be a full-height loop, with simulation of the reactor vessel, wetwell, drywell, GDCS pool (at 1175 scale, rather than 1:25), and IC and PCCS pods and heat exchangers. It is not meant to serve as a full-scope simulation of the reactor and safety systems, but rather as a test of selected components in an integral configuration, primarily to study multidimensional behavior under long-term post-accident conditions. The system will operate at low pressure.

The PANDA tests have been presented by GE as " confirmatory" in nature, rather than " developmental," and therefore not necessary for design certification. The staff's position, however, is that data from these tests are necessary to fulfill--the requirements of 10 CFR 52.47 for analytical model verification and proof of system performance, since previous testing was at much amaller scales and did not represent the current design of the passive heat removal or ECC systems. GE has agreed-to work with PSI to accelerate the PANDA test schedule, to the extent possible, to provide data on system performance before the scheduled date for final design approval.

In addition, Toshiba is modifying the GIRAFFE facility to conform to the current SBWR configuration, with separate IC and PCCS heat exchangers and heat rejection pools, condensate return from the PCCS to the GDCS pool, etc. The staff met with GE on July 20, 1992.to discuss what additional tests should be run in GIRAFFE to provide additional integral system data on long-term (more than-one hour after ECCS initiation) post-accident decay heat removal.

1 i

9  !

The staff expects that an appropriate set of additionel  !

experiments will be identified, and that these tests can be run within the next several months. It should be noted that, ,

although these additional tests will provide valuable systims data in a configuration representative of the current reactor design, the issue of scale discussed previously is still of ,

concern to the staff. The staff still believes that the PANDA l tests, at a substantially larger scale, are necessary in addition ,

to the new tests in the modified GIRAFFE loop.

The staff is concerned that no integral testing appears to be planned that includes all safety systems, key nonsafety systems, and relevant control systems. The potential for systems interactions that could affect.the performance of the passive safety systems must be evaluated, and the staff is continuing to assess the adequacy.of GE's integral test program to address this -

issue. A related and important issue is whether a high-pressure integral test facility is needed to support certification. As noted above, the SBWR does not include any high-pressure-safety-injection systems. While this places a: substantial burden on GE to demonstrate a highly reliable depressurization capability, this issue should be able to be resolved through separate-effects.

tests.

The exclusive reliance on depressurization and-low-pressure safety injection in the SBWR, coupled with a-depressurization logic based on elapsed time after reaching a low vessel level setpoint, appears.to reduce-the potential for interactions at high system pressures that will degrade plant safety performance.

Therefore, based.on information available to the staff at-this time, it does not appear that high-pressure integral testing will be required for design certification of the SBWR.. However, the staff requires that further detailed.information on_ completed and '

planned experimental programs be provided for review to support this position. A primary aspect of this issue is that tests-of the gravity-driven coolant system in-the GIST facility 1were conducted at low pressures.. The transients were simulated  ;

analytically from reactor operating pressure to the pressure at initiation of the test, and the depressurization rate of the experiment.was adjusted to follow the calculated curve at that point. The database used for validation of the models used in

-the high-pressure portion of the simulation, including ,

depressurization rate at-high= pressures through the-DPVs, needs to be justified by the vendor.

The approach for'resolutionHof these issues is the responsibility of the vendor. GE must justify fully the approach chosen for satisfying the-requirements of 10 CFR 52.47 regarding consideration of systems-interactions and adequacy of calculational models. 1

-m. -_ _. _._ _ _ . _ . . . . _ _ _ . _ _ _ . . _ _ __ -.__- ____a,~._

10

8. Eevere Accident Performance Tests In the event of a severe, core-melt accident, both safety-related and non-safety-related systems in the SBWR will be subject to high levels of thermal nnd mechanical loading. In view of the unique features of the plant design, it may be appropriate to require testing of components or systems tLat will be involved in severe accident mitigation or accident management. No specific vendor-sponsored testing related to severe accident performance has been f.dentified at this time for the SBWR besides testing the PCCS using non-condensible gases. Tha staff will be evaluating the applicability of current industry-sponsored research to the SBWR design, as well as determining where additiona) vendor-sponsored testing is appropriate. The staff will utilize the containment performance goal as the most significant measure in evaluating the adequacy of the various severe accident activities. Therefore, it is important to first identify the challenges that may lead to early containment '

failure. In this ~ontext, the staff will use a conditional containment failure probability (CCFP) of 0.1 or a deterministic goal that:

The containment should maintain its role as a reliable leak tight barrier by ensuring that containment stresses do not exceed ASME service level C limits for approximately 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following the onset of core damage, and that following this 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period the containment should continue to provide c barrier against the' uncontrolled release.of fission products.

The containment performance goal is intended to ensure that the containment will perform its function in the face of most severe accident challenges. While the staff has identified both in this paper and in SECY-90-016, " Evolutionary Light Water-Reactor Certification Xssues And Their Relationship To Current Regulatory Requirements," the major challenges to containment-(e.g.,

hydrogen burns, corium interactions with water and containment structures), and the need to provide means for mitigation of these challenges, the containment performance goal acts as a final check to ensure that the design (including its mitigation features) would be adequate if called upon to mitigate a severe accident. The intent of both the CCFP and-the alternative deterministic performance criteria discussed above is to provide this final check as well as defense-in-depth. The philosophy behind the use of the proposed deterministic goal is that adequate time must be provided for fission product decay before allowing a release from the containment to the environment.

Since service level C is applicable only to metal containments, a comparable criterion is needed for the SBWR concrete containment.

The staff is considering the use of ASME factored load category for this criterion.

i i

4 1

l x

=.

u 5

j i

From this vantage point, the first consideration will lan to determine whether.there is a sufficiert understanding?of the phenomena to analyze the. containment response to'these challenges. If there is enough uncertainty, the ongoing test programs will be evaluated and any additional testing will_be identified to develop the necessary confidence ~that-the containment performance goal is met. With this approach in mind, GE abould address the following: _f-

a. Debris Coolability The proposed criteria for debris coolability forLthe SBWR.

is that cavity floor should be sized to provide 0.02m 8/MWt-surface area to promote long-term _ debris coolability and '

assure'long-term containment integrity. The staff believes that debris-coolability objectives should~be coupled to the issue of containment integrity 11n-:that, for example, in the absence of demonstrable complete-debris coolability, an alternate-approach would lus to= ensure that containment overpressurization resulting- from the release of noncondensible-gases due-to core concrete-interaction is consistent'with the containment performance objective. ~

In addition,_ containment basemat and pedestaluwall-erosions should not result in failure of.the containment, again consistent with the containment performance _-objec-tive.

Passive schemes for providing flooding (pro- or: post- .

accident)-of the floor area beneath-the_ vessel should be based on current data regarding debris coolability.- The issue of debris coolability is a.partfof-the existing-research program called Melt Attack and Debris?Coolability Experiment ~(MACE) co-sponsored by the-USNRC, DOE, EPRI, and.several foreigncccuntries.

The. applicability of the MACE data to-the SBWR-design should 1xa evaluated. It is recognized -that there are differences between the experiments and~the:SBWR design, e.g., debris _ depth and composition.- The. determination of the degree of' applicability of theseLexperiments therefore-is-critical in the assessment'of;this design and the~

determination of whether additional testing is needed.

The method of determining applicability could uso supporting _ analyses to aid in maximizing the usefulness of the existing program data to the SBWR. For example, the vendor :should perform a parametric study- to realistically bound the-expected maximum amount of melted core debris

. . 4.

12 that can exit the reactor vessel and that can contribute to concrete ablation and containment pressurization. With this ex-vessel melt mass as the boundary condition, the vendor should assess loadings on the-containment for the first 24 hcurs following a severe accident.

b. Hydrogen Generation and Control The SBWR containment will be inerted to deal with severe accident-generated hydrogen concentrations. The vendor e should justify the long-term reliance on initial inerting or an active inerting system to control hydrogen post-accident. Additionally, because-of the large zirconium inventory, one unresolved issue for the SBWR is the effect of the noncondensible gases on the efficacy of the-isolation and passive containment cooling condensers. The SBWR test program addressing this issue will be closely followed to determine if further testing is necessary,

c. Containment Performance in Severe Accident Natural circulation flow and related mixing processes are key issues for the passive core cooling system (PCCS) and the containment long-term heat removal-system.

Containment codes now available for the analysis of con-tainment phenomena, particularly under severe accident conditions, may not be well-suited for the analysis of natural circulation processes. It is our understanding that the PANDA facility will be used to examine these phenomena for SBWR. . The staff will request GE for further clarification,

d. Fuel Coolant Interaction (FCI).

The vendor's research program does not address.the interaction between core debris and water either within or external to the vessel. The focus of this investigation should be both to explore the credibility of these events and to assure, for those credible events, whether. (1) a coolable configuration will be achieved, (2) FCI steam and hydrogen generation are-accounted =for,-(3) containment response due to dynamic loads from an-FCI are acceptable, and (4) FCl dynamic effectc on~ accident progression areL considered. The impact on the containment of FCI will be assessed consistent with the containment performance objective. Furthermo.re,.the dynamic effect of corium spreading in the presence of a large water-pool has not been completely substantiated by the existing experimental data.

< -o 13

c. Potential for Containment Bypass Unlike the above four issues, which are generic-to the passive plants, the SBh'R has a unique issue associated with the generic concern of containment bypass. The issue is related to the potential of containment bypass through the PCCS. The PCCS transfers heat from the containment atmosphere to an ex-containment water pool. The vendor should address the generic pathways such as isolation failure, containment failure due to wall corrosion, and the testing in place to ensure containment integrity, for example. But, in addition, the vendor should also address the potential for containment bypass resulting from failure of the PCCS heat exchanger tubes following a severe accident, due to elevated containment atmosphere temperatures or carryover of core debris.

ENCLOSURE 3

+

ga n go, UNITED STATES

. 2 ^% NUCLEAR REGULATORY COMMISSION N E ADVISORY COMMITTEE ON REACTOR SAFEGUARDS W ASHINGT ON, D. C. 20555

,8, June 10, 1992 The Honorable Ivan Selin Chairman U.S. Nuclear Regulatory Commission Washington, D.C. 20555

SUBJECT:

TESTING AND ANALYSIS PROGRAMS IN SUPPORT OF THE SIMPLIFIED BOILING WATER RE. ACTOR DESIGN CERTIFICATION During the 385th and 386th meetings of the Advisory Committee on Reactor Safeguards, May 6-9 and June 4-5, 1992, we reviewed the testing and analysis programs in progress and proposed by GE Nuclear Energy (GE) in support of the certification effort for the.

Simplified Boiling Water Renctor (SBWR) passive plant design. Our subcommittee on Thermal Hydraulic Phenomena held meetings to discuss this topic on April 23 and June 2,- 1992. During these meetings, we had tr.e benefit of discussions with representatives of GE and the NRC staff. We also had the benefit of the documents referenced.

GE will use its best-estimate code, TRACG, to evaluate the SBWR thermal hydraulic behavior under accident conditions ranging from ATWS with instabilities to long-term behavior of the Passive containment Cooling System (PCCS). GE representatives presented a very good analysis of processes and phenomena important to accident scenarios postulated for the SBWR. The results were summarf-ed in tables which are to be used by GE to validate the TRACG computer code. However, these same tables appear not to have been used to-guide t' 1 design and operation of the experimental facilities that are to support the code validation process.

The GE experimental program consists of three elements:

1) Laboratory scale experiments to obtain fundamental heat transfer data,

2) Separate effects tests to obtain data for parts of the total system and full-scale components where necessary, and

3) Integral system teste to obtain system data.

4,2m. r) f y K f ! h h ka ? - ..

.*e 2 June 10, 1992 The Honorable Ivan Selin Although we were shown some comparisons of TRACG predictions with data f rom GE's integral system tests (GIST and GIPJ.FFE f acilities) ,

the question of whether or not the facilities can scale the important phenomena was not addressed in either GE's presentation or in the documents supplied to the ACRS by GE. A rigorous scaling analysis is needed if integral system test data alone are to be used to demonstrate that a TRACG calculation is meaningful.

3

- We have some comments about the elements of the GE test plan. The initial conditions for the integral system tests are based on conditions assumed to exist some time after vessel depressurization. These conditions include an initialThe drywell and nitrogen PCCS nitrogen mass fraction of 15 percent.

concentration could be much higher. GE should develop a basis for its choices of initial conditions or broaden its test matrix to higher values of the nitrogen include some tests at much

-d concentration, both in the drywell and in the PCCS.

Separate effects tests to be conducted in the PANTHERS facility will yield the data needed to characterize heat exchanger behavior under a variety of expected conditicns. In particular, GE has agreed to add instrumentation to the individual heat exchanger tubes to obtain local heat transfer data. This will make the GIRAFFE integral system experiments more meaningful. We believe GE has been very responsive to issues raised by both the ACRS and the NRC staff in this regard.

The oscillatory behavior observed in the GIRAFFE integral system tests needs more detailed study to ensure that the suppression pool does not overheat due to steam bypass of the PCCS through the suppression pool top horizontal vents. The steam The flow rate vill be suppression low which could lead to a stratified condition.

pool is not a very of f ective heat sink when this process occurs.

This may well require a separate effects study to obtain data for development of a low steam flow model for the horizontal vent.

Further, review of the GIRAFFE facility instrumentation is needed to ensure that the resulting data will support TRACG model validation.

"'h e SBWR has full pressure isolation condensers (IC) capable of removing 4.5 percent of full power decay heat at full system pressure. The behavior of isolation condensers is well understood and introduces no new processes. GE has indicated that it will collect relevant IC operating data for staff review. The SBWR is automatically depressurized when the vessel water level drops to some prescribed value by a staged opening of squib-type valves.

Further, GE has had a great deal of experience with automatic depressurization and only the squib-type valve itself is of a new design. As a result, we do not believe that full-height, full- {

4

... , I o The Honorable Ivan Selin 3 June 10, 1992 pressure integral system testing is required for certification of the SBWD. design.

The GE program includes conduct of integral system testing at the l PANDA facility located in Switzerland. The NRC staf f would like GE  !

to obtain data from this f acility in time to support its design certification review of the SBWR. To do so, GE would have to l accelerate its schedule by six months. We agree with the NRC staf f f that further integral system testing of the PCCS is needed prior to the final design approval. It has not been demonstrated by GE that existing data obtained from GIRAFFE or GIST testing are sufficient for validation of the TRACG code, nor that the PANDA test facility J will yield.the needed data. A more definitive assessment by GE is j needed; this assessment should include both the scaling rationale i f or the GIRAFFE, GIST, and PANDA f acilities, and a deraonstration of how the effects of test f acility scaling distortion impact the important processes and phenomena outlined by GE in its evaluation of TRACG. As a part of such an effort, it may be possible to show that one can obtain the needed data by some combination of additional separate effects tests and judicious use of the GIRAFFE and GIST facilities.

To summarize, we agr'ee with the NRC staff views that full-height, full-pressure integral system testing is not needed to support the SBWR design certitication. Further, we agree that early integral system testing of the PCCS is essential to meet the present design certification schedu3e. We have not, however, seen evidence that the PANDA facility is adequate to obtain the needed data.

Sincerely O

David A. Ward Chairman

References:

1. Memorandum dated February 26, 1992, for the Commissioners from James M. Taylor, Executive Director for Operations, transmitting Advance copy of proposed Commission paper,

" Evaluation of the General Electric Company's (GE's) Test Program to support Design certification for the Simplified Boiling Water Reactor (SBWR)"

2. Letter dated February 3,1992, from R. C. Mitchell, GE Nuclear Energy, to U.S. Nuclear Regulatory Commission,

Subject:

GE Response to Request for Information on SBWR Testing Program

e ?O The Honorable Ivan Selin 4 June 10, 1992

3. Joint Study Report, " Feature Technology of Simplified BWR (Phase I) GIRAFFE (N.nal Report)," dated November 1990, The Japan Atomic Power Company, et al. (GE Proprietary Information)

4. GE Nuclear Energy, GEFR-00850, " Simplified Boiling Water Reactor (SBWR) Program Gravity-Driven Cooling System (GDCS)

Integrated Systems Test - Final Report," A.F. Billig, dated l October 1989 (Applied Technology Restriction)

5. " ALPHA - The Long Term Passive Decay Heat Removal and Aerosol  !

Retention Program at the Paul Scherrer Institute, i Switzerland," by P. Coddington, et al., Paul Scherrer i Institute, undated

6. Paper from the Proceedings of The International Conference on I Multiphase Flows '91 - Tsukuba, Japan, September 24-27,

" Condensation in a Natural Circulation Loop with Noncondensable Gases Part 1 - Heat Transfer," K. M. Vierow, GE Nuclear Energy, and V. Schrock, University of California

7. GE Draft Report: " Test Specification for IC & PCC Tests,"

undated (GE Proprietary Information)

8. Paper submitted to the Department of Energy, "The Effect of Noncondensable Gases on Steam Condensation Under Forced Convection Conditions," M. Siddique, Ph.D. Thesis -

Massachusetts Institute of Technology, dated January 1992